A newly unveiled Chinese supercomputer has claimed the top position on the authoritative TOP500 ranking of the world’s most powerful computer systems, marking the first time a Chinese system has officially captured the number-one spot in nearly a decade.
Announced Tuesday at the ISC 2026 High-Performance Computing conference, the system, named "LingSheng," achieved a record-breaking sustained processing capability of 2.198 exaflops. The milestone makes it the first supercomputer in the world to officially breach the 2 exaflops performance ceiling on the standardized High-Performance Linpack benchmark, operating at over two quintillion floating-point operations per second.
Developed entirely with domestic Chinese components and architecture, the LingSheng system is deployed at the National Supercomputing Center in Shenzhen. The machine represents a severe break from traditional hybrid acceleration methodologies, utilizing a pure central processing unit design built on China's proprietary LingKun computing platform. The system is powered by 13.79 million independent computing cores consisting of 304-core LX2 processors running at 1.55 gigahertz, all tied together by a specialized custom interconnect called LingQi and running on a domestic Kylin operating system.
Beyond raw computational speed, the platform secured the top ranking on the High-Performance Conjugate Gradients test, an alternative metric designed to evaluate real-world engineering workloads and data access speeds, while placing fourth on the mixed-precision AI benchmarking test. The system draws an estimated 42.2 megawatts of power, yielding an energy efficiency rating of 52.07 gigaflops per watt, which lands it at the fiftieth spot on the accompanying Green500 energy efficiency chart.
The updated global index also featured a significant new deployment from Western Europe. Italian energy conglomerate Eni SpA successfully introduced its new HPC7 system, built by Hewlett Packard Enterprise Co., into the sixth spot on the global chart. That European installation utilizes AMD Zen 4 EPYC processors alongside Instinct MI300A accelerated processing units to deliver 571.50 petaflops of sustained performance, running alongside Eni’s older HPC6 system which currently occupies the eighth global position.